The present invention relates to an electronic fuel supply control method for an automotive engine, or more in particular to a control system equipped with a learning function capable of control under optimum control parameters.
In an internal combustion engine such as a gasoline engine (hereinafter referred to as "the engine"), it is necessary to maintain the amount of fuel supply at a predetermined ratio to the intake, air thereby to keep the air-fuel ratio (A/F) at the right level.
Conventionally, a predetermined air-fuel ratio is obtained by measuring the amount of intake air and by controlling the amount of fuel supply accordingly. However, with this method, satisfactory control is impossible when emission control is taken into consideration.
The trend has thus changed toward the use of an oxygen sensor with zirconia by which the condition of the exhaust gas is detected and the amount of fuel supply is controlled by feedback in what is called the oxygen feedback control system.
In the oxygen feedback control system, a basic fuel supply amount based on the fuel supply amount determined by the above-mentioned amount or flow rate of intake air is compensated for by feedback in a manner to converge the output air-fuel ratio to a predetermined value. As a result, it is possible to drive an automobile always at a predetermined air-fuel ratio even in the case where the air-fuel ratio could not otherwise be kept correctly by controlling the basic fuel supply amount alone.
An example of the engine control system equipped with such an oxygen feedback control device is shown in FIG. 1.
In FIG. 1, reference numeral 1 designates an electronic control system including a microcomputer system, numeral 2 an engine, numeral 3 an oxygen sensor mounted on the exhaust manifold of the engine to determine the output air-fuel ratio from the oxygen concentration of the exhaust, and numeral 4 an injector mounted on the engine intake manifold to inject the fuel.
The electronic control device 1 determines the engine operating conditions in response to the engine intake air flow rate Qa, the engine speed N, the temperature of the cooling water and the battery voltage supplied from sensors not shown, and drives the injector 4 to inject the fuel after further correcting the operating conditions with a signal from the oxygen sensor 3.
The fuel is injected from the injector 4 by periodic interruption in synchronism with the engine revolutions, and therefore, the fuel supply amount is controlled by controlling the fuel injection time of each injection of the injector 4. The injection time Ti is given as EQU Ti=K.multidot.Tp.multidot..alpha..multidot..SIGMA.Ki (1) EQU Tp=Qa/N (2)
where
K: A factor determined by injector PA1 Tp: Basic fuel injection time PA1 .alpha.: Air-fuel ratio control factor PA1 Ki: Various compensation factors PA1 Qa: Intake air flow rate PA1 N: Engine speed (revolutional speed)
As apparent from this equation, the basic fuel injection time Tp is determined by the engine operating conditions, and therefore, it makes up a basic supply amount. In the oxygen feed back method, the control factor .alpha.is changed so that the output of the oxygen sensor 3 alternates between rich and lean states to keep the mean output air-fuel ratio at a predetermined value, that is, a stoichiometric air-fuel ratio (A/F=14.7).
If the basic injection time Tp is kept at the ideal state, the control factor .alpha. pulsates up and down around the level 1.0 and the mean value thereof is 1.0. If the air-fuel ratio based on the basic injection time Tp has changed to the lean side, on the other hand, the control factor .alpha. pulsates around 1.1 in an attempt to correct the situation, while if the air-fuel ratio became 10% richer, the factor .alpha. reciprocates around the level of about 0.9. In each case, the system works to make the output air-fuel ratio an ideal value, and even when the air-fuel ratio given by the basic fuel injection time Tp is displaced from the ideal state, the output airfuel ratio is always kept ideal to prevent the exhaust gas from deteriorating.
In application of this oxygen feedback control method, the response speed thereof has its own practical limit. In the event that the air-fuel ratio based on the basic supply amount undergoes a sudden change, the control operation fails to follow a sudden change of the air-fuel ratio, with the result that the mean value of the output air-fuel ratio deviates from the stoichiometric air-fuel ratio during the transient period before the mean value is converged to a predetermined value, thus deteriorating the exhaust gas. Such a sudden change in the air-fuel ratio based on the basic fuel supply amount is often caused in such cases as when the engine transfers from abrupt acceleration to an engine braked state.
In order to obviate this problem of the oxygen feedback control system, a control method has been suggested and found applications, in which the engine operating conditions are divided into a plurality of regions according to the engine speed or intake air flow rate, and a compensation factor is predetermined for the basic fuel supply amount for each operating region, so that the basic fuel supply amount is corrected by the compensation factor for each engine operating region, thereby keeping the amount of oxygen feedback control substantially unchanged as required against the stoichiometric air-fuel ratio even when the engine operating conditions undergo a change.
In this method, the injection time Ti of the injector 4 is determined by the equation below. EQU Ti=K.multidot.Tp.multidot..alpha..multidot.Kr.multidot..SIGMA.Ki (3)
where Kr is a regional compensation factor.
This method is such that the range of engine speed change and the range of intake air amount change are divided into, say, ten parts respectively, and a total of 100 operating regions are determined by various combinations of the divisions. A regional compensation factor Kr is determined in such a manner as to obtain a stoichiometric air-fuel ratio (=14.7) when the control factor .alpha. is 1.0, that is, when the oxygen feedback control is lacking, for each operating region. The compensation factors thus determined are stored in a memory such as a ROM and are read from time to time during engine operation to calculate the injection time Ti. In this way, it is possible to keep the mean value of the control factor .alpha. substantially at 1.0 to achieve the stoichiometric airfuel ratio and thus the transient deterioration of the exhaust gas which otherwise would occur due to the delayed response of the oxygen feedback control is prevented in any operating region to which the engine operating conditions may change.
The engine control characteristics greatly vary from one engine to another by characteristic variations of the engine or various sensors or actuators used for control thereof.
For this reason, it is substantially useless if a compensation factor Kr required in the regional compensation system which is determined for a standard engine is applied to all other engines. A regional compensation factor Kr must instead be determined independently for each engine and a ROM exclusive to the particular engine is required to store the data. This is, however, impossible to implement as it leads to a lower productivity and increased cost.
The characteristics of the engine, sensors and actuators, on the other hand, are subject to secular variations due to aging, and therefore, the setting of a regional compensation factor during the production process will often become almost meaningless after the lapse of some period of use of the engine.
In view of this, a learning control system has recently been developed. In this system, a non-volatile memory in which data can be written or rewritten is used to store the regional compensation factor Kr, which is sequentially written for each operating region by a "learning" process during engine operation, so that an accurate regional compensation factor Kr is always provide for air-fuel control on the basis of the latest operating results. The basic concept of such a learning control system is disclosed in the Japanese Patents Laid-Open Nos. 20231/79 and 57029/79.
The learning control system eliminates the need of determining a regional compensation factor beforehand, and in case of any change in engine characteristics, etc., enables the regional compensation factor to be corrected by itself from time to time, so that correct control is always possible to prevent the deterioration of the exhaust gas under all operating conditions including the transient period.
In practice, however, this control system fails to produce a sufficient effect, since the engine operations are concentrated in a part of the regions with most of the regional compensation factors left uncorrected.